jdi@Franz.COM (John Irwin) (12/09/89)
Sorry to post this to the net, repeated attempts to send mail to gwm-bugs@mirsa.inria.fr have failed with bizarre mailer-daemon messages. -- First of all, being a Lisp hacker, I'll just say I like gwm a lot! (Although I don't find it easy to use... Hopefully this will change over time as there are more canned examples to look at) Anyway, there is a serious bug in gwm's handling of client reshape that breaks gwm's ICCCM compliance. If a client reshapes and moves a window in one operation, gwm's ConfigureRequestEventHandler first moves the window, sends the client a synthetic ConfigureNotify event, then changes its size, which results in a real ConfigureNotify event. This is not ICCCM compliant. The reshape must result in one and only one ConfigureNotify event. And it should be done in only one ConfigureWindow request, to minimize window "bouncing." Indeed I find a TO_DO note in the function that says "do only one XConfigure..." Well, it turns out that you MUST do only one XConfigure, plus a synthetic ConfigureNotify if only a move was done with no reshape. I'm working on a fix, but it's getting ugly so I thought I'd turn it over to the author... -- John Irwin jdi@franz.com PS: Please add me to the gwm mailing list(s)